Re: name of 'filter' property and corresponding CSSStyleDeclaration property

On Nov 16, 2008, at 7:52 PM, Doug Schepers wrote:

> A possible alternate name, if it comes down to it, might be 'filters'
> (plural), where a list of filters could be specified.  It would
> essentially be an ordered list where the result of the first filter
> would serve as the input for the second filter, as a sort of implicit
> <feMerge>.  This is just brainstorming, though... I haven't thought it
> through thoroughly.  Obviously, another scenario is to rename it  
> with no
> difference in functionality, if there is a suitable name short enough.
>
> However, this is far from optimal.  SVG content is found not only on  
> the
> Web... I've spoken to many people in the Inkscape community for whom  
> the
> Web is simply not a target for their content.  They use SVG as an open
> art medium, an alternative to Illustrator, and they tend to make heavy
> use of filters.

For this reason, we couldn't completely remove 'filter' within one  
spec cycle, so I'm not sure the exercise is worth it, but it's worth  
at least discussing possible solutions.

> Thanks for the context (I hadn't yet followed up on reading the list
> Cameron forwarded the email from, thought I'd guessed it was something
> like that).
>
> Your solution seems fairly sensible, though it's a shame you had to do
> it.  Are the sites "dead" (unmaintained), or is there a possibility  
> you
> could ask them to fix their content?

There were multiple sites affected, enough that we didn't feel  
confident of finding all of them and getting them to change their  
content within a single ship cycle.

> What is the effect on sites that are broken in this way?  Just that  
> they
> use IE-specific opacity?

Objects that should have faded out stayed completely opaque, or vice  
versa. Either way it made features of the site unusable.

> Would another possible solution be for you to
> make a hack that maps that IE-opacity syntax to the standard 'opacity'
> value, such that SVG's 'filter' could be exposed?

That seems like a worse hack than the one we actually did, and more  
fragile, since sites might assume we are IE in other ways besides the  
appropriate syntax for opacity effects.

Regards,
Maciej

Received on Monday, 17 November 2008 06:06:50 UTC